Public | Automated Build

Last pushed: a year ago
Short Description
Docker file for bb_db_stitcher
Full Description

bb_affine_stitcher

Requirements

Good Instruction for installing opencv with opencv_contrib

Usage

Stitching process for points is divided in 2 steps:

  1. Stitching background images
  2. Map Coordinates
  3. Dockerfile

1. Stitching background images

In this step two background images derived in step 1 will be stitched and
the required data for mapping locations (detected in the bb_pipeline) to
the overall coordinate system of one comb side will be saved to a file.

$ bb_fb_stitcher <path_left_img> <path_right_img> <angle_left_rot> <angle_right_rot> <transformation_type> <data_out> -p <preview_img>
  • <path_*_img> - Path to the input images.
  • <angle_*_rot> - initial Rotation angle of both images. Rotation ist measured counterclockwise.
  • <transfromation_type> - defines the transformation used to stitch both images.
  • <data_out> - path for the stitching data, required for further stitching or mapping points. If its a directory, the filename will be 'stitched' of the both input images (Cam_0_2016-09-01T14:20:38.410765Z_ST_Cam_1_2016-09-01T14:16:13.311603Z.npz)
  • -p <preview_img> - creates an preview of the stitched images. (optional)

(See also $ bb_fb_stitcher -h)

Example:

$ bb_fb_stitcher Cam_0_2016-09-01T14\:20\:38.410765Z--2016-09-01T14\:26\:18.257648Z.jpg Cam_1_2016-09-01T14\:16\:13.311603Z--2016-09-01T14\:21\:53.157900Z.jpg 90 -90 3 dir_for_params/ -p preview.jpg

2. Mapping Points

minimal script for mapping points (x,y)

import numpy as np
import fb_stitcher.core as core

camIdx = 0
pts_left_org = np.array([[[1000, 3000], [353, 400], [369, 2703]]]).astype(np.float64)

# initialize Stitcher and load data from step 2
stitcher = core.BB_FeatureBasedStitcher()
stitcher.load_data('Cam_0_2016-09-01T14:20:38.410765Z_ST_Cam_1_2016-09-01T14:16:13.311603Z.npz')

# 'stitch' points
pts_mapped = stitcher.map_cam_coordinates(camIdx, pts_left_org)

print(pts_mapped)

returns [[[ 2990.34596943 3098.03015643] [ 396.12205245 3725.00678215] [ 2694.5516094 3727.65687635]]]

3. Dockerfile

Dockerfile with OpenCV3 and ffmpeg:

$ docker pull gitmirgut/bb_fb_stitcher
$ docker run -it -v   HOST_DIR:/storage gitmirgut/bb_fb_stitcher
Docker Pull Command
Owner
gitmirgut
Source Repository

Comments (0)